Samples and data are analyzed from a longitudinal population study conducted from 1965 to 2007 that allows study of the risk factors and effects of diabetes mellitus. Risk factors for obesity, hypertension, and nephropathy are also studied, along with the relationships of these diseases to diabetes and their effects on development of vascular complications and mortality. The genetics of diabetes is studied by means of family studies and relationships of genetic markers to disease. These findings are reported in DK069028 (Genetic epidemiology of diabetes and obesity) and DK069094 (Genetic epidemiology of diabetic complications). A genome-wide association study implicated variants in the DNER gene in susceptibility to type 2 diabetes. Patterns of DNA methylation (an epigenetic mark) in blood cells were associations with past exposure to diabetes in utero, a strong risk factor for diabetes, suggesting a mechanism for the effect of this risk factor. Knowledge of diabetes risk factors coming from this and other studies led to the hypothesis that type 2 diabetes could be prevented or delayed in adults at high short-term risk. This hypothesis was confirmed in the Diabetes Prevention Program (DPP), a multicenter randomized clinical trial in which many of the participants and investigators in this project participated. We are now in a long-term follow-up phase, the Diabetes Prevention Program Outcomes Study (DPPOS), to assess long-term success with weight loss, reduction in the incidence of diabetes, and effects on diabetes complications. This study is reported in DK075078 (Prevention of type 2 diabetes). We performed another nested case-control study to examine the relationship between serum concentrations of persistent organic pollutants (POPs). The case-control study included 300 persons without diabetes at baseline (median age=24 years); 149 subjects subsequently developed diabetes (cases) and 151 remained non-diabetic during 10 years of follow-up (controls). POPs 37 polychlorinated biphenyls (PCBs) and 9 persistent pesticides (PSTs) were measured in serum collected in 1969-1974. Diabetes was diagnosed clinically or by a 2-hour oral glucose test. Odd ratios for diabetes were computed with logistic regression analysis for total PCBs (sum of the 37 congeners), the individual PCB congeners, and individual PSTs. Three of the POPs, PCB66, PCB151, and PCB195, were weakly but positively associated with development of diabetes. In this nested case-control study, we also measured 25-hydroxyergocalciferol 25(OH)D2, 25-hydroxycholecalciferol 25(OH)D3, and vitamin D binding protein. Half of the study participants (82 cases, 68 controls) were 25(OH)D deficient. Median 25(OH)D3 was 18 ng/ml in cases and 20 ng/ml in controls; 25(OH)D2 was detectable in 12% of cases and 20% of controls. Odds ratio (OR) for incident diabetes in those with detectable vs. undetectable 25(OH)D2 was 0.50 (95% CI, 0.26-0.98) after adjustment for potential confounders. 25(OH) and vitamin D binding protein were not associated with incident diabetes. Detectable 25(OH)D2, derived from dietary sources, was associated with half the rate of incident type 2 diabetes in this population. Whether this finding is attributable to a direct beneficial effect of 25(OH)D2 or to a protective effect of a healthier diet in general remains to be determined. Other findings related to kidney complications of diabetes are reported in DK069062 (Epidemiology, pathophysiology and treatment of diabetic nephropathy).